| name | information-architecture-seo |
| description | Convert a semantic core into crawlable SEO and LLM-friendly information architecture. Use this skill when the user asks for URL structure, page taxonomy, canonical rules, hreflang groups, page roles, pillar/topic/news/article/blog/project separation, multilingual route strategy, or a durable content model such as one short brief plus one longform article. It prevents duplicate public content bodies and hands link-graph work to internal-link-graph-architect. |
Information Architecture SEO
Use this skill after semantic-core-architect and before internal linking, schema, or LLM-readable layers.
Read references/url-architecture.md before producing a full URL map.
Owns
- URL strategy;
- page taxonomy;
- page roles and content ownership;
- canonical model;
- hreflang groups;
- index/noindex policy;
- pillar/topic/news/article/blog/project role separation;
- duplicate-content prevention.
Does Not Own
- semantic core creation;
- final page copy;
- internal link graph details;
- JSON-LD implementation;
- rendered UI design;
- external backlinks.
Workflow
- Read the semantic core, entity-topic map, site constraints, languages, and content model.
- Identify the canonical page role for each cluster/entity/topic.
- Define route patterns before individual URLs.
- Set canonical policy per page role.
- Define hreflang groups for localized equivalents.
- Mark page status:
planned, draft, published, noindex, or deferred.
- Record duplicate-content risks and ownership boundaries.
- Produce
url-map.yaml using assets/url-map.template.yaml.
- Hand off to
internal-link-graph-architect and technical-seo-schema-engineer.
Canonical Rules
- A page can be self-canonical when it has a distinct user/search role and visible content.
- A short news brief and one expanded longform article can both be self-canonical when the brief is the fast news surface and the longform is the durable explanation surface.
- Do not canonicalize a standalone brief to its longform by default.
- Do not create extra public story variants beyond the declared model.
- Topic and pillar pages must summarize and organize; they must not duplicate article bodies.
- Locale alternates use
hreflang; they do not replace canonical self-reference.
Multilingual Rules
- Default language owns
x-default unless the site has a separate language selector page.
- Every localized page must declare its own canonical URL.
- Hreflang groups must contain only true equivalents, not loosely related pages.
- If translation quality is unknown, mark the localized page as
draft or deferred, not published.
Safety And Privacy Boundaries
- Do not expose admin, API, private dashboards, raw analytics, raw IP logs, secret files, or staging URLs as public indexable pages.
- Use
noindex or blocked internal routing for low-value utility pages; do not rely on public navigation hiding.
- Do not weaken authentication, access control, or security headers for crawlability.
- Do not place private source URLs, credentials, tokens, unpublished drafts, or internal notes in
url-map.yaml, sitemap plans, or public LLM-facing artifacts.
Validation
Before marking work complete:
- every page has URL, page type, locale, canonical, index policy, and rationale;
- every language alternate has a coherent hreflang group;
- every story has at most one brief and one longform per language;
- topic pages do not duplicate article bodies;
- final link graph details are left for
internal-link-graph-architect;
- schema implementation is left for
technical-seo-schema-engineer.
Validate skill edits with:
python3 <codex-skills-dir>/senior-skill-architect/scripts/lint_production_skill.py ./skills/information-architecture-seo
python3 ./plans/seo-llm-skill-cluster/scripts/lint_skill_cluster.py .
Forward tests live in evals.json.
Output Shape
Return:
- URL strategy summary.
- Page role table.
- Canonical and hreflang policy.
- Duplicate-content risks.
- Handoff notes for internal linking and technical SEO/schema.
